Joy Buolamwini

Joy Buolamwini coauthored Gender Shades with Timnit Gebru. The paper offers an empirical entry point into evaluating performance across intersecting groups.^[1]

From experience to an audit

In Ford Foundation’s November 13, 2018 filmed profile and published transcript, Buolamwini describes an art project’s face-detection software struggling to detect her face while detecting a white mask. She presents this experience as motivation for investigating the software, not a controlled test on its own.^[2]

She says she started the Algorithmic Justice League after realizing “we were entering the age of automation overconfident and underprepared” (Transcript, introduction). This is her attributed account and public-interest perspective; the video/audio was not independently checked.^[2]

A reading route

Move from her account to the paper’s methods and subgroup results, then NIST’s different evaluation task and the documented arrest case. The route connects personal motivation, measurement and institutional safeguards while preserving each source’s limits. Social Alignment adds the question of whose interests should constrain a system’s use.

Follow the audit over time

Buolamwini coauthored the follow-up study with Inioluwa Deborah Raji.^[3] Read its disclosure process and benchmark-transfer limits alongside Raji’s participant perspective. The route connects measurement with accountability without assuming that any single test establishes acceptable use.